QoS Enhancement in Wireless Multi-Hop Ad Hoc Networks

Abstract

Wireless ad-hoc multi-hop networks such as wireless mesh networks (WMNs) and mobile ad hoc networks (MANETs), have shown a tremendous growth due to the huge demand of mobile users and connectivity strive over the last two decades. Yet, Quality of Service (QoS) of these networks, when traditional protocols such as Transmission Control Protocol (TCP) and Ad hoc On-demand Distance Vector (AODV) are employed, has shown unsatisfactory performance. Throughput unfairness that leads to starvation is a major problem in WMNs backhaul where nodes that are few hops away from the gateway get minimal chance to transmit to the gateway. Additionally, routing in MANETs without considering mobility, residual energy, and congestion impacts QoS severely. This thesis presents designated and tested solutions to enhance the QoS in wireless mesh and mobile ad-hoc networks. These include enhancing throughput with maintaining fairness among nodes in WMNs, and designing efficient routing schemes that able to provide better QoS in terms of packet delivery ratio, delay, network lifetime and energy consumption. Enhanced Adaptive Delayed Acknowledgement Mechanism (EADAM) is proposed to enhance the throughput for all active flows in a static chain topology alongside with maintaining fairness among the flows. This mechanism utilises a mathematical model for calculating appropriate TCP delayed acknowledgement timeout with reference to the number of hops between a source and a destination node called Adaptive Delayed Acknowledgement Mechanism (ADAM). An optimum throughput fairness on a chain topology for a number of active flows has been achieved by implementing a TCP delayed acknowledgement timeout model that allows Paralleled transmission among the individual flows, which leads to a significant throughput enhancement. The proposed mechanism EADAM has been tested in Network Simulator NS2. Then validated by comparing to ADAM. A throughput enhancement ratio of up to 35% has been achieved. Mobility and Energy Aware AODV (MEA_AODV) is a novel route discovery scheme that is mobility and energy aware for AODV in MANETs proposed to improve QoS performance in terms of throughput, Packet Delivery Ratio (PDR), network life time, overall overhead, average delay of an existing scheme called Average Link Stability and Energy Aware (A-LSEA). MEA_AODV relies on two important factors: Link Life Time (LLT) and Residual Energy (RE). Two schemes have been implemented: Selective Energy_ Mobility and Energy Aware AODV (SRE_MEA_AODV) and Selective Link Life Time_ Mobility and Energy Aware AODV (SLLT_MEA_AODV). The two proposed schemes have been tested in NS2, and then evaluated in comparison with AODV and A-LSEA. The evaluation confirms that SRE_MEA_AODV dramatically outperforms A-LSEA and AODV in term of all the QoS aspects. A throughput enhancement of up to 120% compared to AODV and up to 53% compared to A-LSEA achieved with SRE_MEA_AODV. A congestion aware routing scheme has been proposed to provide end‐to‐end guarantees in a route discovery scheme for AODV in MANETs. This scheme improves QoS by controlling the congestion that occurs because of unmanaged node buffer while deciding on forwarding the RREQ packet in the rout discovery process. In this work, the queue length (QL) for the forwarding node is used as a criterion to measure congestion in the node’s buffer. The proposed scheme has been implemented in two schemes called: Firstly, Adaptive Managed Buffer_ route discovery scheme for AODV (AMB_AODV). Secondly, Adaptive Managed Buffer and Selective Remaining Energy route discovery scheme for AODV (AMB_SRE_AODV). The two proposed schemes have been evaluated in comparison with standard AODV. The evaluation shows that PDR has improves to up to 27.9% and 31.2% with AMB_AODV and AMB_SRE_AODV respectively. In addition, throughput enhancement of up to 18.73% and 42.7% achieved with AMB_AODV and AMB_SRE_AODV respectively compared with standard AODV.